Behavioral flexibility is a fundamental cognitive capacity in all mammals. When our environments change, we must be able to adapt our actions. For example, after frequenting a particular restaurant for some time, we recall the precise route and associate this with our favorite dish. However, sometimes circumstances change-the restaurant may change locations-and we must adapt an action (the route we take) in order to receive the same outcome (our favorite dish). A deficit in this capacity is observed in numerous psychiatric disorders including drug addiction, schizophrenia, and obsessive-compulsive disorder and these impairments are generally resistant to treatment. In order to develop improved therapeutics and diagnostic tools, a deeper understanding of the mechanisms underlying impaired behavioral flexibility is necessary. Functional imaging studies in humans as well as lesion studies across species have linked the prefrontal cortex (PFC) with cognitive functions including flexible behavior. However, this brain region does not work in isolation; rather it is part of a larger circuit connecting other structures including the mediodorsal thalamu (MD). Indeed, functional imaging studies show that the MD is important for flexible behavior. However, with imaging studies, causal relationships cannot be determined. Therefore, animal models must be implemented in order to examine the effect of direct manipulation of brain activity. In order to determine whether the MD causally supports flexible behavior, we recently created a mouse model with decreased MD activity and showed that this manipulation led to impairments in behavioral flexibility. However, behavioral flexibility requires multiple cognitive processes. Not only must an organism associate an action with an outcome, but it must also recognize the environmental stimuli that lead to that outcome. Therefore, we next examined the effect of decreasing MD activity during tasks assessing these elementary cognitive processes. We found that the MD plays a role in the formation of action-outcome associations as well as the use of Pavlovian stimuli to shape future actions. These findings establish a role of the MD in specific cognitive processes underlying flexible behavior. However, the MD is part of a larger neural circuit described above. Thus, the roles of specific MD inputs and outputs in the elementary cognitive processes of flexible behavior remain unknown. This proposal implements novel techniques to reversibly manipulate brain activity in the mouse in order to identify the MD neuronal circuitry supporting flexible behavior. The knowledge gained from these studies will aid in the development of new therapeutic strategies.